沙尘暴天气统计预报模型的研究

Research on Statistic Model for Sand and Dust Storm Forecasting

选用气象因子和地表植被因子作为预报因子,其中以归一化植被指数作为植被因子,采用逐步判别分析法对浑善达克地区1988—2000年3—5月份301个大风天气和沙尘天气样本资料进行了统计和分析,建立了沙尘暴天气的多级预报模型,并用2001年3—5月的52个试报样本进行检验,结果表明试报准确率平均为79 4%。

As a meteor, sand and dust storm results from a special atmospheric movement. But weather was not the only factor, the earth's surface had made a significant effect on its forming and development. It's qualitative to predict sand and dust storm, so Fisher's discriminant was used to analyze historical data of wind and sandstorm in the Hunshandake area. Based on the available high-quality surface and meteorological observation data from 1988 to 2001, the predictors were analyzed by stepwise method to build linear equations of different groups.The testing ratio of the equations were 79.4%.